
Learn to build production-ready Salesforce real-time AgentForce agents from scratch using AgentScript, Flows, Apex, and Data Cloud, addressing hallucinations with a structured framework and integrating with external systems like Slack.
Explore how to build a 24/7 hotel assistant with Agent Force, integrating Rumi for bookings, room service, food orders, and checkout in Salesforce using reusable actions and flows.
Design a deterministic agent script framework with an active flow gate to lock context and prevent hallucinations in hotel room booking, using topic selectors, variables, and explicit transitions.
Adds a food booking flow that verifies customers by ID against Salesforce, using an auto-launch validate customer action and an active flow to ensure only verified, checked-in customers proceed.
Build an Apex custom action to process dynamic food orders from a Rumi agent by deserializing JSON of item IDs and quantities, creating a customer-linked food bill and items.
Rumi guides customers from onboarding to checkout, calculating and displaying a line-by-line bill breakdown, instructing payment at the reception, returning keys, and updating Salesforce status to checkout.
Create a public hotel community site and configure a rum i service agent to enable guest chats, set digital experience, and customize welcome and system messages.
Adds a roomie checkout topic that validates the customer, calculates and displays a line-by-line bill, guides payment at the reception, and updates Salesforce to checkout.
Create a knowledge prompt template to enable a clinical trial agent to retrieve trials from a PDF data library, filter by patient age, city, gender, and condition, and present matches.
Publish the logged-in user ID from the community page, populate a pre-chat field, and use an omnichannel flow to stamp the messaging session with the current account for the agent.
Create a flow action that builds a dynamic prompt from the current patient details via session context, then passes it to the clinical trial recommendation prompt and displays results.
Create a backend post enrollment agent separate from the front-end clinical trial agent to manage backend activities with a prompt template and data library.
Create a record-triggered flow on enrollment that invokes the post enrollment agent after a delay, passing participant id and clinical trial name to trigger activities and slack messages.
Welcome to the Salesforce Agentforce Real-Time Project Implementation Course, where you'll learn to build AI agents in real-world scenarios.
The demand for Agentforce is growing rapidly, yet many Salesforce professionals still lack hands-on experience. This course is designed to bridge that gap by guiding you through practical projects that help you build real-time AI agents with confidence.
Whether you're learning Agentforce for the first time or building a POC for your clients, this course offers hands-on experience and the confidence you need to work effectively with Agentforce. you will build Projects that you can confidently explain in an Interview , You’re encouraged to share them with recruiters and hiring managers.
in this course, we’ll build real time projects from scratch:
1. Roomie Agent
In this project, we’ll begin by building our very first AI Agent — Roomie, a virtual assistant for a luxury hotel website. This is a beginner-friendly project that introduces you to the foundational concepts of Agentforce.
You’ll learn how to:
Create your first Agentforce AI Agent
Define reusable actions that power specific tasks like room service and food ordering
Train the agent to ask the right questions at the right time
Leverage context to offer personalized and accurate responses
Integrate the agent with a customer-facing Experience Cloud site, accessible to any hotel guest — no login required
Roomie is designed to assist hotel guests throughout their stay. It can:
Help onboard new customers
Assist with food order placement
Handle room service requests
Support users during the checkout process
You’ll also get hands-on with:
Defining Topics that translate into real agent capabilities
Adding instructions and actions that guide the agent in performing tasks
Building custom actions using Flows and Apex
Creating dependent actions, where the output of one action feeds into the next
By the end of this section, you’ll have built a fully functional AI assistant that simulates a real-world hospitality use case — laying a strong foundation for more advanced projects ahead.
2. Clinical Research Agent (Frontend) And Post Enrollment Agent (Backend)
In this project, we’ll take a major leap and build a much more advanced AI agent — the Clinical Research Agent. This agent is designed for the healthcare and research industry, capable of answering detailed patient queries around clinical research, trial eligibility, and company-specific studies.
You’ll learn how to:
Equip your Agent to Answer User's question by utilizing Agentforce Data library
Build Complex Agent Actions using Knowladge Prompt templates, Apex , Flows and API's
Solve business Use case with Less code and Less Record Creation by giving Agent Access to Unstructured Business Data and Train it to take intelligent Actions.
Work With Context Variable to get all the Share all the details with Agent before the Conversation starts with Customer
Work with Multiple Agents to divide the Task Among them, the Front end agent to talk with Customer and Backend agent to perform backend Analysis and Execution
This agent uses several features of Salesforce Einstein AI and Agentforce, including:
Agentforce & AI Agents
Data Cloud and RAG
Agentforce Data Library
Prompt Builder
Context Variable
Complex Actions using Apex ,flow , API
Invoking Agent with Record Triggered flows
By the end of this section, you’ll be capable of building production-grade AI agents for critical, high-impact industries like healthcare and research
Note:
Code and Prompt Templates Built in Lectures will be provided in the Respective lectures.